In this short article, we will occupy ourselves with the
requirements that can be made of a simulation system for manufacturing
and logistics. The demands may vary according to the exact application.
This article is especially suitable as an aid for the selection of a
simulation system.
When selecting a simulation
system it is important to take a number of matters into account.
Important is the user interface. How easily can the system be used, is
there help for the user and are there checks on input and output?
Regarding
the modeling possibilities: does the system offer enough possibilities
to build a model of the system to be studied? If yes, does the
simulation possibilities meet up to the requirements? Here you can
think of speed, repeating experiments, eliminating start up conditions
and the possibility to intervene during the simulation.
Furthermore,
it is of great importance that you can obtain the output you need, in
standard reports and graphs or in self-definable output. You could
think of information such as utilization, waiting times, throughput
times, numbers produced, delivery reliability, etc.
The
simulation tool will at least have to provide possibilities for
generating arrivals, performing tasks, generating and using chance
numbers and manipulating queues.
Necessary
statistic functions are the possibility of making draws from a number
of probability distributions and the determining of averages, standard
deviations, frequency distributions and reliability intervals.
The
flexibility of a system deserves special attention: is the user limited
to the use of the built-in possibilities or can he add his own logic,
reports, variables, etc? And does the system offer openings with regard
to other packages (exchange of data)? It is also handy when a
simulation system guards the user as much as possible from mistakes by
checking consistency, giving error messages and offering the
possibility of undoing mistakes.
Visualization
and animation are important for conveying ideas to third parties: is
the state of the system in the course of time easy to follow? Use of
icons, dynamic graphical output, visualisation of queues, scaling
possibilities and use of color may prove of invaluable service.
Naturally,
you should always question whether the tool concurs with the desired
application. Is the problem continuous or discrete, stochastic or
deterministic, should the model be detailed or is a high level of
abstraction desired?
Finally, there are a
number of practical considerations to be made, such as: which hardware
is required, what are the costs (software, training, learning time,
etc.) and what kind of support can the supplier offer?